improved gmp compliant approach to manipulate

METHOD Open Access

Improved GMP compliant approach tomanipulate lipoaspirates, to cryopreservestromal vascular fraction, and to expandadipose stem cells in xeno-free mediaFrancesco Agostini1, Francesca Maria Rossi2, Donatella Aldinucci3, Monica Battiston1, Elisabetta Lombardi1,Stefania Zanolin1, Samuele Massarut4, Pier Camillo Parodi5, Alessandro Da Ponte1, Giovanni Tessitori6,Barbara Pivetta6, Cristina Durante1† and Mario Mazzucato1*†

Abstract

Background: The stromal vascular fraction (SVF) derived from adipose tissue contains adipose-derived stromal/stemcells (ASC) and can be used for regenerative applications. Thus, a validated protocol for SVF isolation, freezing, andthawing is required to manage product administration. To comply with Good Manufacturing Practice (GMP), fetalbovine serum (FBS), used to expand ASC in vitro, could be replaced by growth factors from platelet concentrates.

Methods: Throughout each protocol, GMP-compliant reagents and devices were used. SVF cells were isolated fromlipoaspirates by a standardized enzymatic protocol. Cells were cryopreserved in solutions containing differentalbumin or serum and dimethylsulfoxide (DMSO) concentrations. Before and after cryopreservation, we analyzed:cell viability (by Trypan blue); immunophenotype (by flow cytometry); colony-forming unit-fibroblast (CFU-F)formation; and differentiation potential. ASC, seeded at different densities, were expanded in presence of 10% FBSor 5% supernatant rich in growth factors (SRGF) from platelets. The differentiation potential and cell transformationgrade were tested in expanded ASC.

Results: We demonstrated that SVF can be obtained with a consistent yield (about 185 × 103 cells/ml lipoaspirate)and viability (about 82%). Lipoaspirate manipulation after overnight storage at +4 °C reduced cell viability (−11.6%).The relative abundance of ASC (CD34+CD45−CD31–) and endothelial precursors (CD34+CD45−CD31+) in the SVFproduct was about 59% and 42%, respectively. A period of 2 months cryostorage in autologous serum with addedDMSO minimally affected post-thaw SVF cell viability as well as clonogenic and differentiation potentials. Viabilitywas negatively affected when SVF was frozen at a cell concentration below 1.3 × 106 cells/ml. Cell viability was notsignificantly affected after a freezing period of 1 year.Independent of seeding density, ASC cultured in 5% SRGF exhibited higher growth rates when compared with 10%FBS. ASC expanded in both media showed unaltered identity (by flow cytometry) and were exempt from geneticlesions. Both 5% SRGF- and 10% FBS-expanded ASC efficiently differentiated to adipocytes, osteocytes, and chondrocytes.(Continued on next page)

* Correspondence: [email protected]†Equal contributors1Stem Cell Unit, CRO Aviano National Cancer Institute, Aviano, PN, ItalyFull list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Agostini et al. Stem Cell Research & Therapy (2018) 9:130 https://doi.org/10.1186/s13287-018-0886-1

http://crossmark.crossref.org/dialog/?doi=10.1186/s13287-018-0886-1&domain=pdf

http://orcid.org/0000-0001-8319-053X

mailto:[email protected]

http://creativecommons.org/licenses/by/4.0/

http://creativecommons.org/publicdomain/zero/1.0/

(Continued from previous page)

Conclusions: This paper reports a GMP-compliant approach for freezing SVF cells isolated from adipose tissue by astandardized protocol. Moreover, an ASC expansion method in controlled culture conditions and without involvement ofanimal-derived additives was reported.

Keywords: Stromal vascular fraction, Adipose tissue, Freezing protocol, Cell viability, CFU-F, Immunophenotypecharacterization, Adipose stem/stromal stem cells, Growth rate, Differentiation potential, Cell morphology, Karyotype,Anchorage independent growth, Good manufacturing practice, Advanced therapy medicinal product

BackgroundAdipose tissue, beside its role for energy storage, is a knownsource of stromal precursors and stem cells. These cells areenclosed in the so-called stromal vascular fraction (SVF), aheterogeneous population including hematopoietic cells,and adipose-derived stromal/stem cells (ASC). Approachesfor the phenotypic characterization of SVF cells were previ-ously suggested in position papers from the InternationalFederation for Adipose Therapeutics and Science (IFATS)and from the International Society of Cellular Therapy(ISCT), as well as in other publications [1–3]. Such a com-posite and heterogeneous pool of cells can be used for clin-ical applications by virtue of the pro-angiogenetic andimmune modulatory activity exerted by the different cellpopulations [1, 2]. SVF can be extracted from lipoaspiratesby enzymatic digestion, followed by filtration and cell wash-ing [4]. The possibility of storing purified SVF products bycryopreservation and freezing is crucial to optimize thestudy design for clinical applications in humans. SuccessfulSVF cryopreservation using different protocols has beendemonstrated in previously published works [3, 5] but thereis not clear consensus about the appropriate freezing ap-proach that maximally preserves cell viability.In recent years, stromal cells have been utilized in a

growing number of trials for different clinical applications[6]. SVF derived from adipose tissue is an abundant sourceof ASC [7, 8]. IFATS/ISCT proposed three minimalcriteria for the definition of ASC: 1) plastic adherence; 2)expression of CD73, CD90, and CD105, and lack of ex-pression of CD11b, CD14, CD19, CD45, and HLA-DR;and 3) differentiation potential into adipocytes, chondro-cytes, and osteoblasts [9, 10]. Ex-vivo cell expansion mustbe performed to obtain a sufficient number of cells for po-tential use in humans; such “extensive manipulation” issubject to Advanced-Therapy Medicinal Product (ATMP)regulations [11]. ATMPs must be produced in compliancewith current Good Manufacturing Practice (GMP) guide-lines. As well as strict requirements concerning the pro-duction facilities (cell factories, personnel, procedures),GMP guidelines also regulate the quality of reagents anddisposables used in the manufacturing process. Fetalbovine serum (FBS) is a well-known growth supplementfor cell culture, but xeno-carbohydrates and xeno-proteinscontained in FBS may lead to undesired clinical effects

[12–15]. Thus, since 2001, replacement of FBS with suit-able alternatives has been recommended for GMP-gradecell expansion throughout Europe and the United States[16]. The use of human growth factors derived from plate-lets for ex-vivo stem cell expansion is compliant withGMP guidelines [17–20]. In previous publications, growthfactor release from platelets was obtained by repeatedfreeze and thaw cycles [18, 21–24]. In this work, we useda supernatant rich in growth factors (SRGF) produced, aspreviously published [25], through CaCl2 addition to theplatelet apheretic product from healthy donors. SRGF finalbatches used in this work were manufactured by poolingtogether a number of single-donor products that has beenshown to reduce the variability of growth factor concen-trations [26]. The impact on mesenchymal stem cell pro-liferation and differentiation mediated by growth factorsderived from platelets through CaCl2 addition has beenpreviously investigated [17, 27, 28] but it is still not fullycharacterized. Furthermore, there is limited knowledgeabout the influence of SRGF on ASC physiology in cellculture. Noteworthy, in this paper, the SRGF-mediated ef-fect on the ASC growth rate was assessed by seeding cellsat different densities; the effect of plating density on cellphysiology has been previously investigated only on mes-enchymal stem cells derived from murine [29] or human[30–32] bone marrow. In this work, we aimed to definean efficient GMP-compliant method to cryopreserve SVFcells isolated from adipose tissue and to expand ASCusing SRGF as medium additive, starting from a thawedSVF product. In particular, we investigated the impact ofdifferent cryopreserving solutions on viability, immuno-phenotype, clonogenicity, and differentiation capacity ofnucleated cells (NC) or ASC enclosed in SVF. In addition,ASC were expanded at different cell seeding densities in10% FBS- or 5% SRGF-containing media. We determinedthe impact of different culture conditions on the ASCexpansion rate, immunophenotype, and differentiationpotential. GMP-compliant materials, reagents, and de-vices were used.

MethodsDescription of the studyThis study obtained approval from the Ethics Committeeof the CRO Aviano National Cancer Institute (protocol

Agostini et al. Stem Cell Research & Therapy (2018) 9:130 of 16

number: CRO-2016-30), and it was carried out in ac-cordance with the Declaration of Helsinki (2004).In this study, 19 leftover lipoaspirates were processed

for SVF extraction, with 14 lipoaspirates being takenfrom the abdomen and the remaining samples from thethigh or hip/flank. The first six lipoaspirates were imme-diately treated after delivery to the cell-processing la-boratory from the operating theatre, while the otherswere left at +4 °C for 16–18 h (overnight) before ma-nipulation. Cell count and viability assay were performedimmediately after SVF isolation. To optimize the proto-col of such quality control test, the impact of performingred blood cell lysis before cell viability evaluation wastested in paired aliquots of the first four SVF products.Immunophenotypic analysis of SVF cells before cryo-preservation was performed in nine SVF. Cryopreserva-tion was performed using four different solutions. Dueto limited availability of extracted SVF cells, simultan-eous application of all four cryopreservation methods toeach product was not possible; thus, solutions A and Bwere used in 10 out of 19 SVF, solution C in 9 out of 19SVF, and solution D in 5 out of 19 SVF. The compos-ition of each solution is described below. Cell viabilitywas measured again in all SVF products upon thawingafter 2 months of storage. Further investigations wereperformed on five SVF products cryopreserved by solu-tions C and D, in which we evaluated immunopheno-type, colony forming unit-fibroblasts (CFU-F) growth,and differentiation potential. Moreover, three aliquots ofthese SVF products were used to assess viability uponthawing after 1 year of storage. The same five SVF sam-ples that were thawed after 2 months of storage wereplaced in culture to extract and expand the ASC. We de-fined the impact of the cell seeding density on ASCgrowth rate in both 5% SRGF- or 10% FBS-containingmedia. Moreover, we tested the effect of the differentmedia on ASC morphology, immunophenotype, differen-tiation potential, transformation, and karyotype stability.

SVF isolation from adipose tissueLipoaspirates were derived from 19 female breast cancerpatients (age 52.4 ± 1.6 years, body mass index 23.5 ± 2.4kg/m2) who underwent quadrantectomy or totalmastectomy and reconstructive lipotransfer. Fat tissuewas harvested under local anesthesia by standard sterileliposuction techniques as described by Coleman [33, 34].All samples were waste byproducts of surgery. SVFisolation was performed according to a previouslypublished protocol [8] with modifications. Under alaminar flow biological cabinet, the lipoaspirate wastransferred to a plastic bag (Transfer Bag, JMSSingapore PTE LTD, Singapore) and washed three timesby adding warm (37 °C) clinical-grade Ringer Lactate so-lution (S.A.L.F. S.p.A. Laboratorio Farmacologico,

Bergamo, Italy). After phase stratification, the lowerphase was discarded. Thereafter, a solution of collage-nase, 0.15 U/ml final concentration (NB 6 Good Manu-facturing Practice grade, SERVA Electrophoresis GmbH,Heidelberg, Germany) was added to the washed lipoaspi-rate. The product was incubated for 60–70 min at 37 °Cin a certified medical device (Plasmatherm BarkeyGmbH & Co KG, Leopoldshoehe, Germany). The strati-fication (Fig. 1a) was considered to indicate completelipoaspirate digestion. After collagenase neutralizationby the addition of a human albumin solution (Albital200 g/l, Kedrion S.p.A., Lucca, Italy) and product stratifi-cation, the lower phase was recovered and centrifuged at400 × g for 10 min at +4 °C. The cell pellet was washedwith a solution composed of 10% Albital, 10% Anti-coagulant Acid Citrate Dextrose Solution–A (ACD-A;Haemonetics Corporation, Braintree, MA, USA), and2 U/ml heparin (Epsodilave, HOSPIRA ITALIA S.r.l.,Napoli, Italy) in Ringer Lactate solution.

SVF cell count and viability analysisCell counts were performed using a Burker chamber. Toestimate the total NC content, a small aliquot of the SVFfinal product was stained with 5% crystal violet dye(Sigma, St. Louis, MO, USA). SVF cell viability was esti-mated by Trypan blue dye exclusion test; an aliquot of thecell product was added to 0.4% Trypan blue dye (0.2%final concentration) (European Pharmacopeia 7.0, 2011).Only NC with an approximate circular shape were takeninto account and, among these, only markedly coloredcells were considered to be dead cells. In part of the

a

b

Fig. 1 a Distinct stratified phases are clearly evident after 1 h oflipoaspirate treatment with collagenase solution at 37 °C. Stromalvascular fraction cells are contained in the lower clear phase; digestedadipocytes as well as released oily lipids are stratified in the two upperphases. The image is representative of a completely accomplisheddigestion process. b Images of diluted stromal vascular fraction cellscytospun on a glass slide. Cells were differentially stained by May-Grunwald/Giemsa protocol. Scale bar =10 μm


samples, cell counts were performed by three independentand skilled operators to test the reliability of the per-formed evaluations; obtained data were correlated andconsistent (data not shown).Cell viability was calculated as follows: percent NC via-

bility = viable NC / total NC × 100. For cell imaging pur-poses, an aliquot of SVF samples was diluted andcytospun by a standard protocol on a glass slide.Elimination of red blood cells from the specimen can

facilitate microscopic evaluation of NC. Red blood celllysis cannot be performed on the cellular product ad-ministered to the patient since a GMP-grade lysis solu-tion is not available: thus, features of the lysed biologicalspecimen could be different from the product aimed tobe administered to the patient. In this work, we testedthe impact on NC viability results mediated by perform-ing red blood cell lysis immediately before Trypan bluedye exclusion test. On a small aliquot of selected SVFproducts, lysis solution (PharmLyse, BD Biosciences,Becton-Dickinson, San Jose, CA) was added (5 min inwet ice) before cell staining with Trypan blue dye. Inparallel, NC viability was evaluated in the same pairedSVF products by Trypan blue dye staining, without theaddition of lysis solution. In part of the SVF fresh sam-ples, cells were cytospun and differentially stained by theMay-Grunwald/Giemsa protocol. Images were taken bya digital color camera (Motic) (Fig. 1b).

Immunophenotypic characterization of SVF and ASC byflow cytometryAfter lysis of residual erythrocytes in PharmLyse solu-tion (BD Biosciences), cells were washed, maintained inBD Pharmingen Stain Buffer, and labeled with a largepanel of monoclonal antibodies (all from BD Biosci-ences): CD34 APC or PE-Cy7, CD45 APC-Cy7 or FITC,CD31 FITC, CD90 FITC or APC, CD73 PE, CD13 APC,CD44 FITC, CD29 APC, CD166 PE, CD10 PE or APC,HLA I/ABC PE, HLA II/DR FITC, CD106 PE, CD36FITC, CD146 PE, CD235 (Glyα) PE, CD144 PE, CD11bAPC, CD11c APC, CD14 APC, and CD105 PE(Beckman Coulter). Labeling was performed while keep-ing in each tube CD34 (clone 581), CD45, and 7-aminoactinomycin D (7-AAD; to discriminate deadcells). Glyα was used to confirm the exclusion of residualred blood cells from the analysis. In some cases, the SVFsample was labeled as received without lysis or washing,and Hoechst was added to the labeling combination inorder to stain nuclei and confirm the gating strategy(data not shown).Samples were acquired on a BD FACSCanto II flow cyt-

ometer and analyzed by BD Diva software. All experimentswere performed after instrument calibration with CS&Tbeads (BD Biosciences). A multicolor gating strategy ana-lysis was performed, excluding debris (using residual

lymphocytes as an internal standard), and dead cells (basedon 7-AAD), focusing on live intact cells [35]. CD34-positive (+) cells were gated and deeply characterized forexpression of CD31 and subsequently of the othermarkers. Among the CD34-negative(–) cells, the CD45+

fraction was identified as residual white blood cells.Aliquots of ASC, expanded at the seeding density of

1 × 103 cells/cm2 in media with 10% FBS or 5% SRGFadded for different cell passages (see below), weredetached by trypsinization and washed in PharmingenStain Buffer. Cells were labeled with the same panel ofmonoclonal antibodies used for SVF.

Cryopreservation and freezing of SVFCryopreservation was performed by resuspending SVFcell pellets in four different solutions. Solution A wasprepared by adding the following to the standard salinesolution (B. Braun, Melsungen, Germany): Albital (10%),ACD-A (5%), and dimethylsulfoxide (DMSO; 10%;CryoSure-DMSO, Li StarFish, Milan, Italy). Solution Bwas prepared by adding the following to the saline solu-tion (B. Braun): human serum (50%), Albital (5%), ACD-A (2.5%), and DMSO (10%). Solution C and solution Dwere prepared by adding DMSO at 10% and 5%, respect-ively, to pure human serum. Thus, solutions A and Bcontained the saline solution, the same amounts ofDMSO, and different concentrations of human serum,Albital, and ACD-A. The serum-based solutions C andD differed between each other only for the DMSO con-centration, and they did not contain Albital or ACD-A.Solutions A, B, and C contained the same amounts ofDMSO. Final concentrations were reported. Sampleswere frozen overnight at −80 °C in freezing containersdesigned to achieve a cooling rate of −1 °C/min (MrFrosty, Thermo Scientific, Waltham, MA, USA). After-wards, samples were stored in liquid nitrogen vaporphase (−190 °C).

SVF thawingSVF aliquots were thawed at 37 °C for 2 min and, to com-pletely remove DMSO, cells were resuspended in 8 mlserum-free minimum essential medium eagle—alphamodification (α-MEM; Lonza, Basel, Switzerland) contain-ing 100 IU/ml penicillin and 100 mg/ml streptomycin(both from Sigma) and were centrifuged at 260 × g for5 min at room temperature. Serum-free medium was usedto avoid the influence of any medium additive on the fol-lowing cell assays or expansion procedure.

CFU-F assayThe availability of fibroblast progenitors in fresh andthawed SVF aliquots (cryopreserved by methods C andD) was evaluated by CFU-F assay [9, 36, 37]. Cells werethen resuspended and seeded at a final concentration of


5 × 102 NC/cm2 in complete α-MEM. Complete α-MEMwas obtained by adding 100 IU/ml penicillin, 100 μg/mlstreptomycin (both from Sigma), and 10% vol/vol FBS(Lonza). Six-well plates (BD Biosciences, Bedford, MA,USA) were incubated at 37 °C under 5% CO2. After24 h, the supernatant medium containing nonadherentcells was removed and discarded. After a single washwith phosphate-buffered saline (PBS; Lonza, Basel,Switzerland), fresh complete α-MEM containing 10%FBS was replaced on adherent cells. After 14 days, cellswere washed with PBS, fixed with methanol, and stainedby 5% vol/vol crystal violet (both from Sigma). Colonieswere evaluated under an inverted phase-contrast micro-scope (Olympus CKX41, Olympus Italia Srl, Milano,Italy). Aggregates of more than 50 cells were consideredas colonies.

SVF differentiation potential assayTo assess the differentiation potential of freshly extractedor thawed SVF aliquots cryopreserved by methods C andD the following procedure was applied. SVF cells were re-suspended in complete α-MEM containing 10% FBS andthen cells were seeded at 5 × 104 NC/cm2 in standard T25tissue culture flasks (BD Biosciences). After 24 h,nonadherent cells were removed. After a single wash withPBS, α-MEM containing 10% FBS was replaced on adher-ent cells. After 7 days, cells were detached by trypsin-ethylenediaminetetraacetic acid (EDTA; TrypLe Select10X, Life Technologies-Thermo Fisher Scientific,Waltham, MA, USA) and counted using a CellometerAuto M10 (Nexcelom Bioscience, Lawrence, MA, USA).Adipogenic, chondrogenic, and osteogenic differentiationswere achieved using StemMACS AdipoDiff, ChondroDiff,and OsteoDiff media (Miltenyi Biotec GmbH, BergischGladbach, Germany) following the manufacturer’s instruc-tions. After 21 days, adipocytes, chondrocytes (spheroids),and osteocytes were stained by Oil Red-O, Safranin-O,and Alizarin Red (Sigma), respectively. Osteogenic andadipogenic differentiation was analyzed by staining quanti-fication in captured images using MATLAB® (Mathworks,Natick, MA, USA) software. After background subtrac-tion, results were expressed as the percentage coveredarea. Chondrogenic differentiation was quantified measur-ing spheroid size taking advantage of Motic Images Plus 2.0 software®. Spheroid volume was calculated using the for-mula: (width2 × length × 3.14) / 6 [38].

SRGF productionPlatelet concentrates were obtained as previously pub-lished [25] with minor modifications. Briefly, samples ofplatelet concentrates were obtained from washouts ofleukocyte depletion filters intended for disposal, takenfrom platelet apheresis collection kits (HaemoneticsMCS+ System; Haemonetics, Signy-Centre, Switzerland)

after donation from healthy donors. Under a sterile bio-logical safety hood, the filtered apheretic product wastransferred to a sterile tube (BD Biosciences). Platelet ac-tivation was performed by the addition of CaCl2 at afinal concentration of 0.04 M (Monico SPA, Venice,Italy) and by incubation at 40 °C for approximately60 min, i.e., until complete clot formation. Supernatantsof centrifuged samples were stored at −80 °C until ana-lysis. To obtain SRGF batches containing consistentamounts of growth factors we created pools of 16 single-donor SRGF products [26]. The obtained product wasfiltered through a 0.22-μm mesh filter (Merk Millipore,Darmstadt, Germany). Sterility tests were performed bystandard procedures. Two separate batches were utilizedin the present work.

Ex-vivo expansion of ASC from thawed SVF cellsThawed SVF cells were separately plated in standard T25tissue culture flasks (BD Biosciences) with complete α-MEM medium plus 10% vol/vol FBS and in parallel with5% vol/vol SRGF. SRGF was used at 5% as we previouslydemonstrated that further increasing the SRGF concentra-tions failed to confer any relevant advantages in terms ofthe ASC proliferation rate [28]. Cells, seeded at a celldensity of 5 × 104 NC/cm2, were allowed to adhere for24 h. Nonadherent cells were then removed, and freshmedium was added after a single wash with PBS. Upon80–90% confluence, cells (considered to be ASC) weredetached by trypsin-EDTA. Resuspended cells wereseeded (at passage (P)1 and at each following cell passage)at four different densities: 1 × 102 cells/cm2, 1 × 103 cells/cm2, 5 × 103 cells/cm2, and 1 × 104 cells/cm2. At eachpassage, the theoretical cell yield (TCY) was calculated.The following equations were applied: TCY at Pi = n1i

× e[ln(2)/PDTi], where PDTi = (tPi – tPi-1) × 24/PDN atPi. (tPi – tPi-1) × 24 is the time interval (in hours) be-tween consecutive cell passages, and PDN at Pi = 3.32 ×(log n2i – log n1i). n1i is the number of seeded cells andn2i is the number or harvested cells at the selected pas-sage (Pi). When only part of the harvested cells wereseeded, the following equation was applied: TCY at Pi =TCY at Pi–1 × e[ln(2)/PDTi].ASC expanded in α-MEM medium plus 10% vol/vol

FBS or 5% SRGF were considered to be at low passages atP2–P3 [39]. ASC expanded in α-MEM medium plus 10%vol/vol FBS were considered to be at high passages whenbetween P7 and P9 [39]. Otherwise, a posteriori consider-ing the expansion rate of ASC expanded in α-MEMmedium plus 5% vol/vol SRGF, such cells were consideredto be at high passages when between P12 and P14.

Morphometric analysis of expanded ASCImages of high and low passage ASC expanded inα-MEM medium plus 10% vol/vol FBS or with 5% vol/


vol SRGF were taken 24–48 h after seeding at 1 × 103

cells/cm2 by phase-contrast microscopy (Olympus) anddigital color camera (Motic). Morphometric analysis wasperformed taking advantage of Motic Images Plus 2.0software® after appropriate calibration. At least 60 cellsper captured image were manually analyzed. Measuresof the cell major and minor axes as well as cell area werequantified using appropriate software tools following themanufacturer’s protocol.

ASC differentiation potential assayAdipogenic, chondrogenic, and osteogenic differentiationpotential was assayed at high and low passages in ASC ex-panded by complete 10% FBS or 5% SRGF media at theplating density of 1 × 103 cells/cm2. Cells were detachedby trypsinization, and adipogenic, chondrogenic, andosteogenic differentiations were achieved utilizingStemMACS AdipoDiff, ChondroDiff, and OsteoDiff media(Miltenyi Biotec GmbH). After 21 days, osteocytes,adipocytes, and chondrocytes (spheroids) were stained byOil Red-O, Safranin-O, and Alizarin Red (Sigma), respect-ively. Cell images were taken by inverted phase-contrastmicroscope (Olympus) and digital color camera (Motic).Osteogenic and adipogenic differentiation was analyzed bystaining quantification in captured images usingMATLAB® (Mathworks) software. After background sub-traction, results were expressed as the percentage of cov-ered area. Chondrogenic differentiation was quantifiedmeasuring spheroid size taking advantage of Motic ImagesPlus 2.0 software®. Spheroid volume was calculated usingthe formula: (width2 × length × 3.14)/6 [38].

Cell transformation assayASC expanded until high passages in complete 10%FBS or 5% SRGF media at 1 × 103 cells/cm2 weredetached by trypsinization, and resuspended in serum-free α-MEM plus 100 IU/ml penicillin and 100 μg/mlstreptomycin. A total of 4 × 103 cells resuspended in200 μl medium were further diluted with 1 ml of amethylcellulose-based medium (MethoCult H4230,StemCell Technologies, Vancouver, BC, Canada) [40].After repeated resuspension by syringe (BD Biosci-ences) equipped with a blunt-end needle, cells wereplated (in duplicate) in two wells of a 24-well plate (BDBiosciences). After 14 days, colony formation waschecked by inverted microscope. Aggregates containingat least 50 round-shaped cells clearly distinguishablefrom underlying adherent cells were considered as col-onies. HT1080 sarcoma cells (used as a positive con-trol) were expanded in complete α-MEM plus 10% FBS,according to standard cell culture protocols. Imageswere taken by inverted phase-contrast microscope(Olympus) and digital color camera (Motic).

Karyotype analysisASC expanded until high passages in complete 10% FBSor 5% SRGF media at 1 × 103 cells/cm2 were detached bytrypsinization. When a consistent fraction of proliferatinground-shaped cells was evident on phase-contrast micros-copy, karyotype analysis was performed according tostandard procedures [41]. Captured images of metaphasicchromosomes were analyzed by CytoVision® software(Leica, Wetzlar, Germany). At least 20 metaphases wereanalyzed for each cell type.

StatisticsData are presented as mean ± SEM. Paired or unpairedStudent’s t tests were applied to compare mean cell via-bility in the SVF product before freezing (Trypan blueexclusion test vs 7-AAD test; immediate vs delayedlipoaspirate manipulation; with vs without addition ofred blood cell lysis solution). The impact of the differentcryopreserving solutions on SVF cell viability uponthawing was analyzed by one-way analysis of variance(ANOVA) for independent samples. Impact of longerterm cryostorage on SVF samples was analyzed byANOVA for repeated measures. Tukey’s honestly differ-ent significance (HSD) with Bonferroni’s correction waschosen as a post-hoc test. Significance of the differencebetween means in a posteriori identified “High” and“Low” groups was tested by unpaired Student’s t test.Linearity of growth curves was tested calculating R2 as ameasure of goodness of fit of linear regression.Differences between regression coefficients (slopes) ofgrowth curves were tested by the Regression ModelAnalysis Test. Differences in cell morphology wereanalyzed by ANOVA for independent samples withinteraction with Tukey’s HSD with Bonferroni’scorrection as a post-hoc analysis.

ResultsSVF cell characterization before freezingThe mean volume of lipoaspirates was 66.8 ± 4.7 ml.SVF mean cell yield evaluated at the end of the extrac-tion protocol (before freezing) was 185.4 ± 19.3 × 103

NC/ml lipoaspirate. We also evaluated cell viability byTrypan blue dye exclusion test in the fresh SVF and, asdisplayed in Fig. 2a, NC viability measured in a subset offresh SVF aliquots treated by red blood cell lysis solutionwas significantly (p < 0.01) lowered when compared withmatched untreated SVF aliquots (77.1 ± 3.9% vs 66.3 ± 4.9%). Figure 2b shows that the mean NC viability(measured without red blood cell lysis) in SVF extractedfrom lipoaspirates stored for 16–18 h at +4 °C wassignificantly (p < 0.01) decreased when compared withimmediately processed samples (82.2 ± 1.8 vs 70.6 ± 2.1).Freshly isolated SVF samples were processed for


cytofluorimetric analysis. The mean NC viabilityevaluated through 7-AAD was 72.2 ± 7.6%.Immunophenotypic characterization was performed

adopting a multicolor strategy that allowed identificationof different vital cell populations. In particular, as shownin Fig. 3a, we identified in the CD34+CD45− population(58.1 ± 7.6% of NC) a CD34++CD31−SSChigh subset (ASC,putative adipose-derived stromal cells; 58.8 ± 16.6% ofCD34+ cells) and a CD34+CD31+SSClow subset (EPC,putative endothelial progenitor cells; 43.2 ± 16.6% ofCD34+ cells).The phenotype of CD34+ cells, and in particular of

ASC, was then characterized in detail with a large panelof antibodies, as reported in Table 1 (part A) and in partshown in Fig. 3b. ASC were brightly positive for CD90and CD73, positive for CD13, CD44, CD10, and HLA I/ABC, dimly positive for CD105, CD29, CD166, CD106,and CD146, and negative for CD36, CD144, CD11c,CD11b, CD14, Glyα, and HLA II/DR.As shown in Fig. 3, when compared with ASC, the EPC

subset CD34+CD31+SSClow expressed CD90 and CD73less brightly, but it was characterized by higher expressionof CD105 and CD146. EPC cells were CD13−. Whencompared with ASC, the EPC population expressedhigher levels of CD144, CD36 and HLA II/DR, CD44 andCD10, while the expression level of the remainingmarkers was not different from ASC (data not shown).Among the CD34− cells, we distinguished a CD34−CD45+

cell population of putative hematopoietic origin (32.5 ± 14.6% of NC) and a CD31−CD45−CD146+ small subsetconsidered as pericytes (Fig. 3b). Overnight lipoaspiratestorage at +4 °C before SVF isolation did not affect theexpression pattern of NC immunophenotype markers.

Impact of freezing on SVF cellsCells were cryopreserved at the mean final concentrationof 1.43 ± 0.7 × 106 NC/ml in the different cryopreservingsolutions (A, B, C, and D; see methods). The impact of

cryopreservation on SVF mononuclear cell viabilitymeasured (upon thawing) by Trypan blue dye exclusiontest is shown in Fig. 4a. Statistical analysis showed thatpost-thaw NC viability in SVF aliquots frozen for 2 monthswith both solutions A and B was significantly (p < 0.01)lower when compared to the viability measured in SVFbefore freezing. Otherwise, cell cryopreservation by solu-tions C and D fully prevented the loss of cell viability afterfreezing. In addition, no significant differences in post-thaw NC viability could be observed when comparing so-lution C with solution D. Since NC viability in SVF sam-ples cryopreserved by solutions A and B was markedlylower (−34.6 ± 4.9%) when compared with samples cryo-preserved by solutions C and D, solutions A and B werenot considered as eligible for clinical applications. Wethen investigated the impact of NC concentration in thefreezing mixture on post-thaw SVF cell viability. Cell via-bility data obtained in thawed SVF samples after 2 monthsof storage with solutions C and D were grouped togetherand a posteriori stratified according to NC concentrationmeasured in each sample upon freezing (i.e., after resus-pension in cryopreserving solution). Two groups wereidentified across a defined threshold of 1.3 × 106 NC/ml:the group (n = 10) showing significantly higher NCconcentration (1.95 ± 0.19 × 106 NC/ml, p > 0.05) wasdefined as High, and the other group (n = 4) was definedas Low (0.52 ± 0.12 × 106 NC/ml). As displayed in Fig. 4b,mean post-thaw cell viability measured in the High group(69.0 ± 3.2%) was significantly (p < 0.05) higher than in theLow group (37.5 ± 7.7%). In parallel, we aimed to evalu-ate thawed NC viability changes after a prolongedcryopreservation period. NC viability measured after1 year of storage in solutions C and D was not sig-nificantly different when compared with viability re-sults measured after 2 months of freezing (Fig. 4c).The phenotype evaluated after short- and long-termfreezing (data not shown) was superimposable on thatobtained in fresh samples (Table 1).

a b

Fig. 2 a Impact of red blood cell (RBC) lysis on percent nucleated cell (NC) viability measures in SVF. b Impact of overnight lipoaspirate storage at+4 °C on percent NC viability measured by Trypan blue dye exclusion test (without red blood cell lysis) in fresh SVF samples. *p < 0.01, vs withoutRBC lysis (Student’s t test for paired data); §p < 0.05, vs immediate manipulation (Student’s t test for unpaired data)


Functional characterization of SVF cellsThawed SVF cells were seeded on a tissue cultureplastic surface and adherent putative ASC wereanalyzed. As shown in Fig. 5, we demonstrated adipo-genic, osteogenic, and chondrogenic differentiation ofadherent ASC cells (P1) before SVF freezing and aftercryopreservation with solutions C and D. Representa-tive images of adipogenic, osteogenic, and chondro-genic differentiation assays are reported in Fig. 5; asfurther evidenced by analytical quantification, the twocryopreservation methods did not affect the cell dif-ferentiation potential. Representative images of CFU-Fcolonies obtained in fresh SVF samples as well as inSVF samples frozen by cryopreserving solutions Cand D are displayed in Fig. 5.

Ex vivo ASC expansionWe expanded ASC extracted from SVF thawed productscomparing, at different cell seeding densities, the effect

of media supplementation with 5% SRGF and 10% FBSon the ASC proliferation rate. Figure 6a shows, in alogarithmic scale, changes in TCY over time (days) mea-sured by culturing cells in the presence of 5% SRGF or10% FBS at different plating densities. When cultured inthe presence of 10% FBS or 5% SRGF, ASC seeded at1 × 103 cells/cm2, 5 × 103 cells/cm2, and 1 × 104 cells/cm2

underwent several passages in culture without evidenceof growth rate reduction. When compared with 10%FBS, 5% SRGF induced an overall higher ASCproliferation rate independent of cell seeding density(Fig. 6a). Considering the same cell culture medium,slopes of growth curves (linear, R2 = 0.98 ± 0.01) werenot significantly affected by cell seeding densities at 1 ×103 cells/cm2, 5 × 103 cells/cm2, or 1 × 104 cells/cm2.Otherwise, when seeded at 1 × 102 cells/cm2 both in 10%FBS and 5% SRGF media, growth arrest rapidlyoccurred. In particular, ASC cultured in 10% FBSunderwent growth arrest at P1 in ASC derived from 2

a

b

Fig. 3 Representative flow cytometry immunophenotype analysis of SVF cells evaluated before freezing. a Gating strategy identifying three mainpopulations in the SVF: CD34+CD31−CD45− subset (ASC, red), CD45−CD34+CD31+ subset (EPC, green), and CD34–CD45+ subset (hematopoietic cells,blue). Dead cells (7AAD+) were excluded. b A detailed CD34+ cell characterization, showing expression of CD13, CD105, CD73, and CD90 in ASC andEPC. Pericytes were identified as CD34−CD45−CD31−CD146+ population (in violet). Lymphocytes are showed as reference (dark blue)


out of 5 patients, at P3 in 2 out of 5, and P6 in the lastcase. Similarly, when seeded at 1 × 102 cells/cm2 inpresence of 5% SRGF, growth arrest occurred at P1 inASC derived from 1 out of 5 patients, at P4 in 3 cases,and at P6 in the last case. At the seeding density of 1 ×102 cells/cm2, growth in isolated cell aggregates wasfrequently observed (data not shown). Figure 6b showsrepresentative images of cultured ASC in 10% FBS- or in5% SRGF-supplemented medium at P0, at low passages,and at high passages. Morphometric analysis of plastic-adherent ASC expanded in 5% SRGF or 10% FBS con-taining media is reported in Fig. 6c; independent of cellpassage number, ASC expanded in 5% SRGF mediumwere smaller than ASC expanded in 10% FBS medium(p < 0.001). The cell culture method and the cell passagenumber significantly interacted (p = 0.002) to modulate

adherent cell dimensions. Post-hoc analysis demon-strated that the cell area of ASC cultured in 5% SRGFmedium was greater at high passages when comparedwith P0 and low passage cells. Still, independent ofcell passage, ASC expanded in 5% SRGF mediumwere more elongated than ASC expanded in 10% FBSmedium (p < 0.001). The effect of cell culture and cell pas-sage number significantly interacted (p < 0.001) to modu-late the adherent cell shape. Post-hoc statistical analysisdemonstrated that, when compared with ASC grown in10% FBS medium, ASC expanded in 5% SRGF mediummaintained a greater axis ratio, especially at high passage.The immunophenotype of ASC expanded in 10% FBS or5% SRGF medium was analyzed both at early and latepassages. The panel of analyzed surface markers was sub-stantially the same as that adopted to evaluate NC in SVFsamples, including markers suggested by IFATS/ISCT [9].As shown in Table 1 (part B), cells expressed CD90,CD73, CD13, CD105, CD29, and CD166 at high intensity,were positive for CD10, CD44, and HLA-ABC, and werenegative for CD34, CD45, CD31, HLA-DR, CD106, CD36,CD146, CD235, CD144, CD11b, CD11c, and CD14. More-over, CD13, CD105, CD29, and CD166 were overex-pressed in expanded ASC when compared withnonexpanded ASC in SVF. CD34 expression was lostalong with expansion, as expected. The expression level ofthe selected surface markers was stable from cell passageP3, and was maintained until late passages, with no sub-stantial differences between cells grown with 10% FBS or5% SRGF (Table 1, part B). Representative images of thedifferentiation potential of ASC expanded in 10% FBS or5% SRGF at low and high cell passages are reported in Fig.7. As also evidenced by analytical quantification, theadipogenic, osteogenic, and chondrogenic differentiationpotential was shown to be not significantly affected whencomparing ASC expanded in 10% FBS and in 5% SRGFcontaining media, both at high and low passages.Chromosome number and structure were analyzed in ex-panded ASC at high passage, and Fig. 8a shows examplesof the obtained karyograms. In all analyzed samples, atleast 20 metaphases were analyzed, and no clonal or re-current chromosomal alterations could be identified.Figure 8b shows representative images of colony for-

mation assays in methylcellulose medium performed onhigh-passage ASC cultured in 5% SRGF- and 10% FBS-containing media. In all the performed assays, ASC col-ony formation on methylcellulose failed to be seen.

DiscussionIn the present study, we describe a method to extract SVFcells from lipoaspirates derived from breast cancer pa-tients who underwent quadrantectomy or total mastec-tomy and reconstructive lipotransfer. A total of 19lipoaspirates were processed to reliably set up and define

Table 1 Expression level of surface markers analyzed in adiposetissue-derived stem cells (ASC) present in the stromal vascularfraction (SVF) and in expanded ASC

Markers Part A Part B

ASC in SVF Expanded ASC

Fresh Thawed 10% FBS 5% SRGF

CD34 ++ ++ – –

CD45 – – – –

CD31 – – – –

CD90 ++ ++ ++ ++

CD73 ++ ++ ++ ++

CD13 + + ++ ++

CD105 +− +− ++ ++

CD44 + + + +

CD29 +− +− ++ ++

CD166 +− +− ++ ++

CD10 + + + +

HLA I, ABC + + +− +−

HLA II, DR – – – –

CD106 +− +− – –

CD36 – – – –

CD146 +− +− – –

CD235 – – – –

CD144 – – – –

CD11b – – – –

CD11c – – – –

CD14 – – – –

Part A reports the expression levels of selected markers in the 34++ 31−SSChigh

ASC contained in fresh or thawed SVF samplesPart B reports the expression levels of the same markers in expanded ASC inthe presence of 10% fetal bovine serum (FBS) or 5% supernatant rich in growthfactors (SRGF) and alpha-minimal essential medium (α-MEM)Expression levels are described as + and – considering unlabeled cells as reference(see also Fig. 3): ++, bright expression; +, positive expression; +−, dim expression


the isolation protocol. Furthermore, we identified a safemethod to cryopreserve and freeze SVF cells with minimalimpact on cell viability or clonogenic and differentiationpotential. Due to the limited amounts of extracted cellsfrom each lipoaspirate, we could apply the different cryo-preservation approaches only to subgroups of SVF sam-ples. Moreover, we investigated the impact on the ASCproliferation rate, identity, differentiation potential, andcell stability mediated by SRGF, considered as a GMP-compliant medium additive for the expansion of ASC toobtain an ATMP. For practical and technical reasons, in-vestigations regarding in vitro expanded ASC were limitedto five SVF specimens as a starting product.The mean quantity of NC extracted from lipoaspirates

as well as the variability of the cell extraction yield (coef-ficient of variation, 42.5%) was in agreement with previ-ous publications [3, 42]. Interindividual differencesbetween lipoaspirate donors could explain the observedvariability [3, 42]. To assay the SVF cell viability we uti-lized the Trypan blue dye exclusion manual test, as sug-gested by European Pharmacopeia [43]. Viability resultspublished in previous papers were higher [3, 42] or

comparable [5] when matched to our present data. Cellviability data were sufficiently consistent since the coeffi-cient of variation between collected results was 13.1%.The presence of dead cells in fresh SVF products can beascribed to mechanical adipose tissue disruption duringthe liposuction procedure, as well as to cell separation,washing, and concentration steps along with theisolation protocol [5]. In accordance with previous data[44], we showed that overnight storage of lipoaspiratesat +4 °C exerts a mild detrimental effect on fresh SVFcell viability. Evaluation of cell viability is a quality con-trol test within the manufacturing process and, thus, thebiological specimen should at best represent the productcondition as administered to the patient. We avoidedred blood cells lysis, frequently used to facilitate evalu-ation of cell viability, as such a procedure is not nor-mally performed on the final product aimed at patientadministration. Due to extensive adipose tissue washing,the occurrence of residual red blood cells in the finalproduct was minimized. Moreover, our results suggestedthat the lysis procedure could cause an underestimationof fresh SVF cell viability.

a

c

b

Fig. 4 a Impact of different cryopreservation solutions on percent nucleated cell (NC) viability in thawed SVF products after 2 months of storagein liquid nitrogen. A, B, C and D are different cryopreservation solutions: solution A (10% Albital, 5% ACD-A, 10% DMSO, 75% saline solution), solution B(50% human serum, 5% Albital, 2.5% ACD-A, 10% DMSO, 32.5% saline solution), solution C (90% human serum, 10% DMSO), solution D (95% humanserum, 5% DMSO). When compared with solutions C and D, the viability of NC stored for 2 months by solutions A and B was significantly lower. *p < 0.01,vs C and D; NSa, not significantly different vs C and vs Pre (one-way ANOVA for independent samples). b Impact of total NC concentration on SVF freezingon post-thaw cell viability. In the high NC concentration group (High), cell viability measured after thawing was significantly higher than in the low concen-tration (Low) group. §p < 0.05, vs High Group (Student’s t test for unpaired data). c Impact of longer term cryostorage on SVF samples. NC viabilitymeasured after 1 year of freezing was not significantly different when compared with results obtained after 2 months storage. *p < 0.01, vs Pre (one-wayANOVA for independent samples). NSb, not significantly different vs 2 months (one-way ANOVA for repeated measures). Tukey’s honestlydifferent significance with Bonferroni’s correction as post-hoc test


The characterization of the SVF cells was performedthrough a multiparametric immunophenotype analysisbased on flow cytometry. Unfortunately, no unique singlemarkers have been found so far; therefore, the analysis of acombination of markers is necessary to identify a cell sub-set sharing the same function and phenotypic signature.Results derived from the large panel of analyzed antigenswere consistent with the minimal criteria proposed in theIFATS/ISCT position paper [9] and with previous reports[45–47]. In particular, as shown in Fig. 3 and Table 1, weidentified a cell population CD34++CD31−CD45−,previously defined [9, 48] as ASC with mesenchymalphenotype, and a cell subset CD34+CD31+CD45− that wecan bona fide define as EPC. In addition, we could identifya CD34−CD31−CD45−CD146+ small population that canbe putatively classified as pericytes [48]. The availability ofsuch cell populations in fresh and thawed SVF confirmsthe possibility of applying this product in regenerativemedicine applications since ASC, pericytes, andendothelial cells can synergistically cooperate to induce theformation of new blood vessels in an optimal regenerativemicroenvironment [2]. In this study, SVF aliquots werefrozen using different cryopreservation solutions; NC

viability was minimally affected in SVF samplescryopreserved with solutions containing pure serum andDMSO at both final concentrations of 5% and 10%. Thus,as previously demonstrated [3, 5], DMSO concentrationsin pure serum can be reduced to 5% to correctlycryopreserve SVF cells. Under such conditions, cellviability in the frozen product was shown to be stable forat least 1 year. In addition, we have further demonstrated[3, 5] that SVF cells can be frozen in serum with 5% (and10%) DMSO without affecting post-thaw CFU-F growth aswell as ASC differentiation capacity. Reducing DMSO givesa significant advantage for clinical utilization of the thawedproduct since DMSO is known to be cytotoxic. Our ap-proach can be considered as potentially compliant withGMP guidelines since animal-derived components werenot used, and certified products or disposables were usedfor the manufacturing process. Moreover, we showed thatpost-thaw cell viability can be strongly reduced when freez-ing SVF at a cell concentration lower than 1.3 × 106 NC/ml (putative threshold roughly corresponding to 1.0 × 106

viable NC/ml). The definition of limited cell concentrationfor optimal freezing is crucially important for futureclinical applications.

Vol.

(mm

3 )

0.00.20.40.60.81.0

Solu�on C

Adipogenic

Osteogenic

Chondrogenic

Uns�mulated

Solu�on D

S�mulated

Uns�mulated S�mulated

BEFORE SVF FREEZING

DIFFERENTIATION POTENTIAL

2 MONTHS SVF FREEZING

DIFFERENTIATION POTENTIAL

CFU-F GROWTHSolu�on C

CFU-F GROWTH

Solu�on D

C.A.

(%)

C.A.

(%)

C.A.

(%)

C.A.

(%)

Vol.

(mm

3 )

0.00.20.40.60.81.0

048

121620

012345

012345

048

121620

Fig. 5 Representative images of osteogenic, adipogenic, and chondrogenic differentiation assays as well as colony forming unit-fibroblast (CFU-F)assays performed on stromal vascular fraction (SVF) cells before freezing or after 2 months of storage in the presence of cryopreservation solutions(Sol.) C and D. Differentiation was induced (Stim.) by the addition of commercially available osteogenic, adipogenic, and chondrogenic differentiationmedia to cells at passage P1. Unstimulated cells (Unst.; control) were cultured with 10% FBS medium. Within the chondrogenesis assay, spheroids failedto be obtained from unstimulated cells. Differentiated cells as well as unstimulated samples were stained with Alizarin Red, Oil Red-O, and Safranin-O todetect osteocytes, adipocytes, and chondrocytes, respectively. The differentiation degree was quantified by image analysis of cell staining (adipogenesisand osteogenesis) or by morphometric analysis of spheroids (chondrogenesis); results are reported in histograms. Sample storage in the presence ofsolutions C and D did not significantly affect cell differentiation potential. Scale bar = 100 μm. C.A., covered area; Vol., volume


In the second part of this work, we aimed to optimizethe ASC expansion protocol. ASC expansion ismandatory for cell therapy applications in humans. Asan ATMP, expanded ASC must be produced in compli-ance with current GMP guidelines. FBS is progressivelyreplaced in cell cultures by growth factors derived fromhuman alternative sources [16] since xeno-carbohydratesand xeno-proteins may lead to undesired clinical effects[12–15]. Mesenchymal stem cells were previously ex-panded in vitro in a serum-free medium with a mixtureof commercially available growth factors; nevertheless,under such conditions, the expression of selected surfacemarkers was shown to be affected [49, 50]. Moreover,utilization of a coating substrate allowing cell adhesionis often required and the cost of media and reagents isconsiderably high. Utilization of human platelet-derivedgrowth factors is compliant with GMP guidelines [17–20, 25]. In previous publications, growth factors werederived by repeated freeze and thaw cycles [18, 21–24].Cells grown under such conditions showed morphology,

as well as proliferation, immunomodulation, and differ-entiation potential, comparable with cells grown in FBS-containing media [18, 51, 52]. In this study, we usedSRGF as standardized medium additive to stimulate ex-vivo ASC proliferation [25, 28]. Knowledge on the im-pact of such a medium additive on ASC physiology incell culture is limited. Ancillary product standardizationis suggested by GMP guidelines [21, 53] and, as we pre-viously demonstrated, pooling together 16 single-donorproducts allowed a satisfactory batch-to-batchconsistency [26]. Our results regarding cell growth kinet-ics demonstrated that SRGF dramatically increased theASC proliferation rate when compared with FBS; this ef-fect was demonstrated considering changes in cell yieldat different cell passages. The expansion rate of bonemarrow mesenchymal stem cells in the presence ofgrowth factors derived from CaCl2-activated plateletswas previously shown to be higher when compared withstandard FBS [17, 28]. In a recent work, we demon-strated that, when compared with FBS and platelet lysate

a

bc

Fig. 6 a Growth curves (logarithmic scale) of adipose tissue-derived stem cells (ASC) seeded at different cell densities (from 1 × 102 to 1 × 104 cells/cm2) in culture medium containing 10% fetal bovine serum (FBS) or 5% supernatant rich in growth factors (SRGF). The presence of SRGF in the cellculture induced a significantly higher growth rate when compared with FBS. b Plastic-adhering cells at P0 and after short-term (Low passage) orlonger-term (High passage) expansion in the presence of 10% FBS or 5% SRGF in the cell culture medium. Scale bars = 100 μm. c Cell morphometricanalysis. ASC expanded in 5% SRGF medium were smaller than those expanded in 10% FBS medium. The cell area of ASC cultured in 5% SRGF mediumwas greater at high passage when compared with P0 and low-passage cells. ASC expanded in 5% SRGF medium were more elongated than ASCexpanded in 10% FBS medium during all culture phases. Linearity of growth curves was tested by calculating R2 as a measure of goodness of fit of linearregression. Differences between regression coefficients (slopes) of growth curves were tested by a Regression Model Analysis Test. *p < 0.01, vs FBS. Errorbars of ASC growth curves could not be graphically reported in the diagram (logarithmic scale of y axis); the coefficient of variation regarding each plotted(mean) value of theoretical cell yield was below 10%. Connectors in c link significantly different means (p < 0.001, ANOVA for independent samples withinteraction with Tukey’s HSD with Bonferroni’s correction as post-hoc analysis). MEM, minimum essential medium; SVF, stromal vascular fraction


as medium additives, SRGF induced the highest prolifer-ation rate also in bone marrow mesenchymal stem cells[54]. In this study, we tested ASC growth kinetics at dif-ferent seeding densities and, independent of cell mediacomposition, we showed that seeding cells at 1 × 103

cells/cm2 can represent an optimal choice to expandASC, reducing repeated exposures to trypsin whileobtaining a satisfactory final cell yield. However, in aprevious publication, ASC displayed changes in geneexpression profiles in relation to cell seeding density:proliferation-related genes were highly expressed in cellsexpanded at low density, whereas genes regulatingchemotaxis or differentiation properties were highlyexpressed in ASC expanded at high density (5 × 103

cells/cm2) [55]. We demonstrated that ASC, rapidlyexpanded in the presence of 5% SRGF at 1 × 103 cells/cm2, were characterized by a satisfactory capacity todifferentiate into osteoblasts, chondrocytes, andadipocytes. Cell morphology was previously associatedwith proliferation rate and differentiation potential and,in particular, small and spindle-shaped cells were recog-nized as rapidly dividing cells, while bigger flatter oneswere considered as slowly replicating cells [56, 57].

Considering our cell morphology analysis, we demon-strated that, when compared with FBS, the ASC ex-panded in SRGF-containing medium were smaller atearly passages and generally more elongated. These re-sults are in accordance with our data demonstrating ahigher ASC proliferation rate in the presence of SRGF,even at extended cell passages.To confirm the identity of ASC, a panel of surface

markers was analyzed considering the ISCT recommen-dations [10]. The cell immunophenotype of ASCexpanded in 10% FBS or 5% SRGF medium was ana-lyzed both at low and high passages, and the obtainedexpression pattern of surface markers was substantiallyin line with previous reports [1, 9, 10]. Moreover, fromthe early passages, a pure and stable cell populationsharing the same surface marker expression profile wasdetected for ASC expanded both in 10% FBS and 5%SRGF medium. Furthermore, the antigen expressionlevel was not differently affected throughout the expan-sion process. Moreover, no difference was demon-strated when considering different cryopreservingsolutions (namely, solutions C and D) used to store theSVF product (data not shown). As expected [1], when

Fig. 7 Representative images obtained from osteogenic, adipogenic, and chondrogenic differentiation assays performed on ASC after short-termor longer-term expansion at 1 × 103 cells/cm2 in the presence of 10% fetal bovine serum (FBS) or 5% supernatant rich in growth factors (SRGF) inthe cell culture medium. The differentiation degree was quantified by image analysis of cell staining (adipogenesis and osteogenesis) or by morphometricanalysis of spheroids (chondrogenesis); results are reported in histograms. The differentiation potential was shown to be not significantly affected whencomparing ASC expanded in 10% FBS and in 5% SRGF-containing media, both at high and low passages. Scale bar = 100 μm. C.A., covered Area; MEM,minimum essential medium; Vol., volume; Unst., unstimulated


compared to nonexpanded ASC contained in the SVF,CD34 expression was blunted in expanded ASC andthe expression level of CD13, CD105, CD29, andCD166 was increased (Table 1, part B). Under our ex-perimental conditions, ASC did not show the expres-sion of CD36; this marker was also reported to benegative in bone marrow mesenchymal stromal/stemcells [45, 58]. Within the process of expansion, ASCcould potentially develop genetic instabilities [59], andchromosomal alterations may lead to apoptosis or celldeath, and also to cell transformation [60]. In thisstudy, we analyzed the cell karyotype and no genetic le-sions in ASC expanded at high passages in the presenceof both 5% SRGF and 10% FBS were seen. Transformedcells normally gain the capacity to grow underanchorage-independent conditions [61]. ASC expandedin both SRGF- and FCS-containing media failed toform cell colonies on methylcellulose [40]. In a previ-ous work, ASC expanded in a xeno-free medium didnot display features of tumor transformation despite ahigh proliferation rate [62]. Thus, considering our re-sults, we can suggest that ASC expanded in 5% SRGFmedium can be considered as genetically stable. Never-theless, only further controlled clinical studies will shed

light on the long-term tumor formation risk secondaryto cell therapy treatments with expanded ASC.

ConclusionsIn conclusion, in this study we identified a safe and GMP-compliant protocol to extract, freeze, and thaw NCcontained in SVF from adipose tissue. Freezing SVF ali-quots at the appropriate cell concentration with a cryopre-serving solution containing low (5%) DMSOconcentration in pure serum failed to affect cell viabilityafter short- and medium-term storage. The availability ofan appropriate extraction, freezing, and thawing protocolis very important to manage the timing of product admin-istration to patients. Afterwards, we provided a better de-scription of the influence of SRGF as medium additive onthe fundamental features of expanded ASC in vitro.Immunophenotype characterization and functional prop-erties of such rapidly proliferating cells were in accordancewith published guidelines [9, 10] and expanded ASC werenot transformed. In this way, we defined a safe method toobtain, by a GMP-compliant protocol, expanded ASCfrom thawed SVF. This whole approach was set up inorder to be easily “translated” within an authorized GMPmanufacturing facility for the production of ATMPs.

a

b

Fig. 8 a Representative karyotypes of adipose tissue-derived stem cells (ASC) expanded at high passages in 10% fetal bovine serum (FBS)- or 5%supernatant rich in growth factors (SRGF)-containing medium. At least 20 metaphases were analyzed and no clonal or recurrent chromosomal al-terations could be identified. b Displays images taken from colony formation assays in methylcellulose medium performed on high-passage ASCcultured in 5% SRGF- or 10% FBS-containing medium. ASC expanded utilizing both cell culture media failed to display colony formation. HT1080fibrosarcoma cells were used as positive control (c+). Scale bar = 100 μm. MEM, minimum essential medium


AbbreviationsASC: Adipose-derived stromal/stem cells; ATMP: Advanced-Therapy MedicinalProduct; CFU-F: Colony forming unit-fibroblasts; DMSO: Dimethylsulfoxide;FBS: Fetal bovine serum; GMP: Good Manufacturing Practice; MEM: Minimumessential medium; NC: Nucleated cells; PBS: Phosphate-buffered saline;SRGF: Supernatant rich in growth factors; SVF: Stromal vascular fraction;TCY: Theoretical cell yield

AcknowledgementsWe acknowledge the nurses and the technical staff of the Stem Cell Unitand of the Breast Surgery Unit of the CRO-IRCCS, Aviano, Italy.

FundingThe project was founded by Italian Ministry of Health (RF 2010 2317993).

Availability of data and materialsThe datasets used and/or analyzed during the current study are availablefrom the corresponding author upon reasonable request.

Authors’ contributionsFA: study design, experiment/analysis performance, protocol optimization,data analysis, manuscript writing and revision; FMR: study design, experiment/analysis performance, data analysis, manuscript revision; MB: experiment/analysisperformance, data analysis, manuscript revision; EL, SZ, BP, and GT: experiment/analysis performance, manuscript revision; SM and PCP: study design, lipoaspiratecollection, manuscript revision; DA, ADP, and CD: study design, manuscriptrevision; MM: scientific coordination, study design, manuscript revision andapproval. All authors read and approved the final manuscript.

Ethics approval and consent to participateThe study was approved by the Ethics Committee of the CRO Aviano NationalCancer Institute (protocol number: CRO-2016-30), and it was performed inaccordance with the Declaration of Helsinki (2004). Signed informed consentwas collected from patients.

Competing interestsThe authors declare that they have no competing interests.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims in publishedmaps and institutional affiliations.

Author details1Stem Cell Unit, CRO Aviano National Cancer Institute, Aviano, PN, Italy.2Clinical-Experimental Onco-Hematology Unit, CRO Aviano National CancerInstitute, Aviano, PN, Italy. 3Molecular Oncology Unit, CRO Aviano NationalCancer Institute, Aviano, PN, Italy. 4Breast Surgery Unit; CRO Aviano NationalCancer Institute, Aviano, PN, Italy. 5Department of Plastic and ReconstructiveSurgery, University of Udine, Udine, Italy. 6Cytogenetic Unit, AAS 5 FriuliOccidentale, “S. Maria degli Angeli” Hospital, Pordenone, Italy.

Received: 5 February 2018 Revised: 5 April 2018Accepted: 23 April 2018

References1. Guo J, Nguyen A, Banyard DA, Fadavi D, Toranto JD, Wirth GA, et al. Stromal

vascular fraction: a regenerative reality? Part 2: mechanisms of regenerativeaction. J Plast Reconstr Aesthet Surg. 2015;69:180–8.

2. Nguyen A, Guo J, Banyard DA, Fadavi D, Toranto JD, Wirth GA, et al. Stromalvascular fraction: a regenerative reality? Part 1: current concepts and reviewof the literature. J Plast Reconstr Aesthet Surg. 2015;69:170–9.

3. Minonzio G, Corazza M, Mariotta L, Gola M, Zanzi M, Gandolfi E, et al. Frozenadipose-derived mesenchymal stem cells maintain high capability to growand differentiate. Cryobiology. 2014;69:211–6.

4. Carvalho PP, Gimble JM, Dias IR, Gomes ME, Reis RL. Xenofree enzymaticproducts for the isolation of human adipose-derived stromal/stem cells.Tissue Eng Part C Methods. 2013;19:473–8.

5. Thirumala S, Gimble JM, Devireddy RV. Cryopreservation of stromal vascularfraction of adipose tissue in a serum-free freezing medium. J Tissue EngRegen Med. 2010;4:224–32.

6. Squillaro T, Peluso G, Galderisi U. Clinical trials with mesenchymal stem cells:an update. Cell Transpl. 2016;25:829–48.

7. Zhu M, Heydarkhan-Hagvall S, Hedrick M, Benhaim P, Zuk P. Manualisolation of adipose-derived stem cells from human lipoaspirates. J Vis Exp.2013;79:e50585.

8. Zuk PA, Zhu M, Mizuno H, Huang J, Futrell JW, Katz AJ, et al. Multilineagecells from human adipose tissue: implications for cell-based therapies.Tissue Eng. 2001;7:211–28.

9. Bourin P, Bunnell BA, Casteilla L, Dominici M, Katz AJ, March KL, et al.Stromal cells from the adipose tissue-derived stromal vascular fraction andculture expanded adipose tissue-derived stromal/stem cells: a jointstatement of the International Federation for Adipose Therapeutics andScience (IFATS) and the International Society for Cellular Therapy (ISCT).Cytotherapy. 2013;15:641–8.

10. Dominici M, Le BK, Mueller I, Slaper-Cortenbach I, Marini F, Krause D, et al.Minimal criteria for defining multipotent mesenchymal stromal cells. TheInternational Society for Cellular Therapy position statement. Cytotherapy.2006;8:315–7.

11. Sensebe L, Gadelorge M, Fleury-Cappellesso S. Production of mesenchymalstromal/stem cells according to good manufacturing practices: a review.Stem Cell Res Ther. 2013;4:66.

12. Abbott A, Cyranoski D. Biologists seek to head off future sources ofinfection. Nature. 2003;423:3.

13. Hill AF, Desbruslais M, Joiner S, Sidle KC, Gowland I, Collinge J, et al. Thesame prion strain causes vCJD and BSE. Nature. 1997;389:448–50. 526

14. Mackensen A, Drager R, Schlesier M, Mertelsmann R, Lindemann A. Presenceof IgE antibodies to bovine serum albumin in a patient developinganaphylaxis after vaccination with human peptide-pulsed dendritic cells.Cancer Immunol Immunother. 2000;49:152–6.

15. Spees JL, Gregory CA, Singh H, Tucker HA, Peister A, Lynch PJ, et al.Internalized antigens must be removed to prepare hypoimmunogenicmesenchymal stem cells for cell and gene therapy. Mol Ther. 2004;9:747–56.

16. Committeee for medicinal products for human use (CHMP). Note forguidance on the use of bovine serum in the manufacture of humanbiological products. Eur Med Agency. 2012;1:1–8.

17. Amable PR, Teixeira MV, Carias RB, Granjeiro JM, Borojevic R. Mesenchymalstromal cell proliferation, gene expression and protein production in humanplatelet-rich plasma-supplemented media. PLoS One. 2014;9:e104662.

18. Schallmoser K, Bartmann C, Rohde E, Reinisch A, Kashofer K, Stadelmeyer E,et al. Human platelet lysate can replace fetal bovine serum for clinical-scaleexpansion of functional mesenchymal stromal cells. Transfusion (Paris). 2007;47:1436–46.

19. Schallmoser K, Rohde E, Bartmann C, Obenauf AC, Reinisch A, Strunk D.Platelet-derived growth factors for GMP-compliant propagation ofmesenchymal stromal cells. Biomed Mater Eng. 2009;19:271–6.

20. Shih DT, Chen JC, Chen WY, Kuo YP, Su CY, Burnouf T. Expansion of adiposetissue mesenchymal stromal progenitors in serum-free mediumsupplemented with virally inactivated allogeneic human platelet lysate.Transfusion (Paris). 2011;51:770–8.

21. Bernardi M, Albiero E, Alghisi A, Chieregato K, Lievore C, Madeo D, et al.Production of human platelet lysate by use of ultrasound for ex vivoexpansion of human bone marrow-derived mesenchymal stromal cells.Cytotherapy. 2013;15:920–9.

22. Capelli C, Pedrini O, Valgardsdottir R, Da RF, Golay J, Introna M. Clinicalgrade expansion of MSCs. Immunol Lett. 2015;168:222–7.

23. Hemeda H, Giebel B, Wagner W. Evaluation of human platelet lysate versusfetal bovine serum for culture of mesenchymal stromal cells. Cytotherapy.2014;16:170–80.

24. Shih DT, Burnouf T. Preparation, quality criteria, and properties of humanblood platelet lysate supplements for ex vivo stem cell expansion. NBiotechnol. 2015;32:199–211.

25. Durante C, Agostini F, Abbruzzese L, Toffola RT, Zanolin S, Suine C, etal. Growth factor release from platelet concentrates: analyticquantification and characterization for clinical applications. Vox Sang.2013;105:129–36.

26. Agostini F, Polesel J, Battiston M, Lombardi E, Zanolin S, Da PA, et al.Standardization of platelet releasate products for clinical applications in celltherapy: a mathematical approach. J Transl Med. 2017;15:107.

27. Parsons P, Butcher A, Hesselden K, Ellis K, Maughan J, Milner R, et al.Platelet-rich concentrate supports human mesenchymal stem cellproliferation, bone morphogenetic protein-2 messenger RNA expression,


alkaline phosphatase activity, and bone formation in vitro: a mode of actionto enhance bone repair. J Orthop Trauma. 2008;22:595–604.

28. Borghese C, Agostini F, Durante C, Colombatti A, Mazzucato M, Aldinucci D.Clinical-grade quality platelet-rich plasma releasate (PRP-R/SRGF) fromCaCl2-activated platelet concentrates promoted expansion of mesenchymalstromal cells. Vox Sang. 2016;111:197–205.

29. Neuhuber B, Swanger SA, Howard L, Mackay A, Fischer I. Effects of platingdensity and culture time on bone marrow stromal cell characteristics. ExpHematol. 2008;36:1176–85.

30. Bartmann C, Rohde E, Schallmoser K, Purstner P, Lanzer G, Linkesch W, et al.Two steps to functional mesenchymal stromal cells for clinical application.Transfusion (Paris). 2007;47:1426–35.

31. Both SK, van der Muijsenberg AJ, van Blitterswijk CA, de Boer J, de Bruijn JD.A rapid and efficient method for expansion of human mesenchymal stemcells. Tissue Eng. 2007;13:3–9.

32. Colter DC, Class R, DiGirolamo CM, Prockop DJ. Rapid expansion ofrecycling stem cells in cultures of plastic-adherent cells from human bonemarrow. Proc Natl Acad Sci USA. 2000;97:3213–8.

33. Coleman SR. Structural fat grafting: more than a permanent filler. PlastReconstr Surg. 2006;118:108S–20S.

34. Coleman SR, Saboeiro AP. Fat grafting to the breast revisited: safety andefficacy. Plast Reconstr Surg. 2007;119:775–85.

35. Zembruski NCL, Stache V, Haefeli WE, Weiss J. 7-Aminoactinomycin D forapoptosis staining in flow cytometry. Anal Biochem. 2012;429:79–81.

36. Astori G, Vignati F, Bardelli S, Tubio M, Gola M, Albertini V, et al. “In vitro”and multicolor phenotypic characterization of cell subpopulations identifiedin fresh human adipose tissue stromal vascular fraction and in the derivedmesenchymal stem cells. J Transl Med. 2007;5:55–65.

37. Castro-Malaspina H, Gay RE, Resnick G, Kapoor N, Meyers P, Chiarieri D, et al.Characterization of human bone marrow fibroblast colony-forming cells(CFU-F) and their progeny. Blood. 1980;56:289–301.

38. Dolznig H, Rupp C, Puri C, Haslinger C, Schweifer N, Wieser E, et al.Modeling colon adenocarcinomas in vitro a 3D co-culture system inducescancer-relevant pathways upon tumor cell and stromal fibroblastinteraction. Am J Pathol. 2011;179:487–501.

39. Wu P-K, Wang J-Y, Chen C-F, Chao K-Y, Chang M-C, Chen W-M, et al. Earlypassage mesenchymal stem cells display decreased radiosensitivity andincreased DNA repair activity. Stem Cells Transl Med. 2017;6:1504–14.

40. Capelli C, Gotti E, Morigi M, Rota C, Weng L, Dazzi F, et al. Minimallymanipulated whole human umbilical cord is a rich source of clinical-gradehuman mesenchymal stromal cells expanded in human platelet lysate.Cytotherapy. 2011;13:786–801.

41. Caspersson T, Zech L, Johansson C, Modest EJ. Identification of humanchromosomes by DNA-binding fluorescent agents. Chromosoma. 1970;30:215–27.

42. SundarRaj S, Deshmukh A, Priya N, Krishnan VS, Cherat M, Majumdar AS.Development of a system and method for automated isolation of stromalvascular fraction from adipose tissue lipoaspirate. Stem Cells Int. 2015;2015:109353–62.

43. European Pharmacopeia 7.0. Nucleated Cell count and viability. EurPharmacop 70. 2011;2.7(29):233–4.

44. Eom YW, Lee JE, Yang MS, Jang IK, Kim HE, Lee DH, et al. Rapid isolation ofadipose tissue-derived stem cells by the storage of lipoaspirates. YonseiMed J. 2011;52:999–1007.

45. Baer PC, Kuci S, Krause M, Kuci Z, Zielen S, Geiger H, et al. Comprehensivephenotypic characterization of human adipose-derived stromal/stem cells andtheir subsets by a high throughput technology. Stem Cells Dev. 2013;22:330–9.

46. Yoshimura K, Shigeura T, Matsumoto D, Sato T, Takaki Y, Aiba-Kojima E,et al. Characterization of freshly isolated and cultured cells derived fromthe fatty and fluid portions of liposuction aspirates. J Cell Physiol. 2006;208:64–76.

47. Zimmerlin L, Donnenberg VS, Pfeifer ME, Meyer EM, Peault B, Rubin JP, etal. Stromal vascular progenitors in adult human adipose tissue. Cytom A.2010;77:22–30.

48. Zimmerlin L, Donnenberg VS, Rubin JP, Donnenberg AD. Mesenchymalmarkers on human adipose stem/progenitor cells. Cytom A. 2013;83:134–40.

49. Al-Saqi SH, Saliem M, Asikainen S, Quezada HC, Ekblad A, Hovatta O, et al.Defined serum-free media for in vitro expansion of adipose-derivedmesenchymal stem cells. Cytotherapy. 2014;16:915–26.

50. Mark P, Kleinsorge M, Gaebel R, Lux CA, Toelk A, Pittermann E, et al. Humanmesenchymal stem cells display reduced expression of CD105 after culturein serum-free medium. Stem Cells Int. 2013;2013:698076.

51. Bernardo ME, Avanzini MA, Perotti C, Cometa AM, Moretta A, Lenta E, et al.Optimization of in vitro expansion of human multipotent mesenchymalstromal cells for cell-therapy approaches: further insights in the search for afetal calf serum substitute. JCell Physiol. 2007;211:121–30.

52. Capelli C, Domenghini M, Borleri G, Bellavita P, Poma R, Carobbio A, et al.Human platelet lysate allows expansion and clinical grade production ofmesenchymal stromal cells from small samples of bone marrow aspirates ormarrow filter washouts. Bone Marrow Transpl. 2007;40:785–91.

53. Araki J, Jona M, Eto H, Aoi N, Kato H, Suga H, et al. Optimized preparationmethod of platelet-concentrated plasma and noncoagulating platelet-derived factor concentrates: maximization of platelet concentration andremoval of fibrinogen. Tissue Eng Part C Methods. 2012;18:176–85.

54. Bernardi M, Agostini F, Chieregato K, Amati E, Durante C, Rassu M, et al. Theproduction method affects the efficacy of platelet derivatives to expandmesenchymal stromal cells in vitro. J Transl Med. 2017;15:90.

55. Kim DS, Lee MW, Yoo KH, Lee TH, Kim HJ, Jang IK, et al. Gene expressionprofiles of human adipose tissue-derived mesenchymal stem cells aremodified by cell culture density. PLoS One. 2014;9:e83363.

56. Docheva D, Padula D, Popov C, Mutschler W, Clausen-Schaumann H,Schieker M. Researching into the cellular shape, volume and elasticity ofmesenchymal stem cells, osteoblasts and osteosarcoma cells by atomicforce microscopy. J Cell Mol Med. 2008;12:537–52.

57. Sekiya I, Larson BL, Smith JR, Pochampally R, Cui JG, Prockop DJ. Expansionof human adult stem cells from bone marrow stroma: conditions thatmaximize the yields of early progenitors and evaluate their quality. StemCells. 2002;20:530–41.

58. Smiler D, Soltan M, Albitar M. Toward the identification of mesenchymalstem cells in bone marrow and peripheral blood for bone regeneration.Implant Dent. 2008;17:236–47.

59. Holzwarth C, Vaegler M, Gieseke F, Pfister SM, Handgretinger R, Kerst G, etal. Low physiologic oxygen tensions reduce proliferation and differentiationof human multipotent mesenchymal stromal cells. BMC Cell Biol. 2010;11:11.

60. Roos WP, Kaina B. DNA damage-induced cell death by apoptosis. TrendsMol Med. 2006;12:440–50.

61. Ochi N, Tanasanvimon S, Matsuo Y, Tong Z, Sung B, Aggarwal BB, et al. Proteinkinase D1 promotes anchorage-independent growth, invasion, and angiogenesisby human pancreatic cancer cells. J Cell Physiol. 2011;226:1074–81.

62. Paula AC, Martins TM, Zonari A, Frade SP, Angelo PC, Gomes DA, et al.Human adipose tissue-derived stem cells cultured in xeno-free culturecondition enhance c-MYC expression increasing proliferation but bypassingspontaneous cell transformation. Stem Cell Res Ther. 2015;6:76.


improved gmp compliant approach to manipulate

Documents